The Agent Skills Directory

[SAFE]: The skill consists primarily of natural language instructions for an AI agent to perform technical synthesis. It does not include any scripts (.sh, .py, .js) or executable binaries.
[SAFE]: Metadata and provenance files (UPSTREAM.md, PATCHES.md) correctly identify the source as a known developer repository (mattpocock/skills) and indicate the files are pinned and unmodified snapshots.
[INDIRECT_PROMPT_INJECTION]: The skill processes conversation context to generate specs. While this is an ingestion surface for untrusted data, the skill is safe as it lacks dangerous capabilities (like shell execution or arbitrary file writes) that could be exploited by an injection in the conversation. It only produces a text-based specification following a defined template.
[COMMAND_EXECUTION]: The skill mentions /setup-matt-pocock-skills and refers to issue tracker publishing, but these are handled by the agent's existing toolset (e.g., GitHub CLI) rather than custom, risky scripts embedded in the skill.

to-spec